HOmics: Integrates two omic data through hierarchical modeling

In lnonell/HOmics: Hierarchical omics

Description

Integrates two omic data through hierarchical modeling

Usage

HOmics(data.matrix, cond, z.matrix, covar.matrix = NULL,
  agg.matrix = NULL, seed = NULL, cores = 1, n.adapt = 1000,
  n.chains = 3, n.iter = 2000, ...)

Arguments

data.matrix	matrix with features as rownames and samples as columns
cond	response variable, usually a numerical factor with two levels representing the conditions to compare. If cond is a numerical vector (continuous response), a hiearchical linear regression model will be fit instead of the default hierarchical logistic regression model
z.matrix	matrix with prior information related to features, with rownames the features and columns the samples
covar.matrix	vector or matrix of continuous covariates, with samples as rownames (in the same order as cond) and covariates as columns. Default = NULL
agg.matrix	matrix with features as rownames and columns corresponding to the groups according to some feature aggregation criteria, 0 for non pertenance. If not specified, analysis will be performed by feature, univariate. Default = NULL
seed	numerical seed for the use of function set.seed in the generation of the model, for reproducibility
cores	cores in case of parallelization. Default = 1 (no parallelization)
n.adapt	number of iterations for the adaptative phase of the hierarchical model. Default = 1000
n.iter	number of iteractions for the burn in phase or sampling of the hierarchical model. Default = 2000
n.chain	number of chains of the hierarchical model. Default = 3

Value

an object of class HOmics

Examples

to be built

lnonell/HOmics documentation built on July 23, 2019, 1:10 a.m.